Conventionally, a natural language classifier is trained by use of machine learning and rules engines to resolve disambiguation with intents associated with same utterances of a query.